Improving Word Embedding Factorization for Compression using Distilled Nonlinear Neural Decomposition.
Vasileios LioutasAhmad RashidKrtin KumarMd. Akmal HaidarMehdi RezagholizadehPublished in: EMNLP (Findings) (2020)
Keyphrases
- recurrent neural network model
- network architecture
- neural model
- data compression
- neural network
- image compression
- co occurrence
- nonlinear manifold learning
- compression scheme
- text compression
- n gram
- singular value decomposition
- compression algorithm
- compression ratio
- manifold embedding
- nonlinear dynamics
- multiscale
- nonlinear dimensionality reduction
- information hiding
- pairwise
- natural language processing
- low rank
- high order